Assessment of Cyberattack Detection-Isolation Algorithm for CAV Platoons Using SUMO

Abstract

A Connected Autonomous Vehicle (CAV) platoon in an evolving real-world driving environment relies strongly on accurate vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communication for its safe and efficient operation. However, a cyberattack on this communication network can corrupt the appropriate control actions, tamper with system measurement, and drive the platoon to unsafe or undesired conditions. As a first step toward practicable resilience against such V2V-V2I attacks, in this paper, we implemented a unified V2V-V2I cyberattack detection scheme and a V2I isolation scheme for a CAV platoon under changing driving conditions in Simulation of Urban MObility (SUMO). The implemented algorithm utilizes vehicle-specific residual generators that are designed based on analytical disturbance-to-state stability, robustness, and sensitivity performance constraints. Our case studies include two driving scenarios where highway driving is simulated using the Next-Generation Simulation (NGSIM) data and urban driving follows the benchmark EPA Urban Dynamometer Driving Schedule (UDDS). The results validate the applicability of the algorithm to ensure CAV cybersecurity and demonstrate the promising potential for practical test-bed implementation in the future.